Good morning everyone. We'll get started with the basic science debrief of the liver meeting 2022. I am so grateful for the leaders of the AASLD and the organizers of the educational programming to have been invited to share this debrief with you. The opportunity to consolidate some of this year's basic science is quite an honor. We came, we saw, and soon we will go, but not before having been exposed to some of the world's best scientific discoveries in hepatology. I am Ritanya Kaur and I am the chief of GI at the University of Washington. I'm a hepatologist and physician scientist who studies basic mechanisms of lipid droplet regulation. Here are my disclosures. I have three objectives for all of you today. The first is that we provide an overview of the basic science landscape at the liver meeting. The second is to identify the major scientific areas of investigation that were represented here, and the third is to summarize key basic science studies that were presented here. In my view, the basic science landscape at the liver meeting 2022 reminds me of my own state of Washington's Mount Rainier. Expansive, majestic, and spectacular. Basic science was represented in all of the major fora here, from the post-grad course to the poster debriefs and sessions, SIG sessions, symposia, workshops, state of talks, parallels, and plenaries. In total, investigations were presented at nearly 60 fora, and our basic science work comprised a large percentage of the nearly 5,000 abstracts that were presented here. We kicked off the meeting with the presidential choice lecture on xenotransplantation, where Dr. Muhedin took us on a journey of the first successfully transplanted, genetically modified, porcine-to-human cardiac transplant. It really was a Herculean effort of sorts that started with a patient with end-stage cardiac disease who had no other opportunities for remedy. The story included an outsized porcine heart, and basic science knowledge from decades of preclinical studies extending all the way back to Dr. Starzl's pioneership. It also included the application of a novel ex vivo heart perfusion system, and these findings were recently published in the New England Journal of Medicine, including the potential mechanisms of the patient's ultimate heart failure and demise at 60 days. Dr. Utrecht then took us through the high Zimmerman talk, where he reminded us that DILI is, well, complicated, especially immune-mediated DILI. He reminded us that it starts with the exposure of the drug and their creation of neo-antigens that are recognized by our immune cells, that both our adaptive and innate immune cells are involved, but it really is dependent on the activity of CD8 positive T cells for some of the action that leads to IDILI. In this case, he shared with us, depletion of CD8 T cells can protect against Ectovacone-induced liver injury. From the weekend forward, we were exposed to a number of abstracts, posters, symposia, and unfortunately, I don't have opportunity to share all of them with you here, but I did take the liberty to do a sort of unofficial review, if you will, of each of the talks and abstracts, and I used a word cloud strategy to view the major themes. The larger the font, the more common the topic. Represented here are the relative frequency of the topics explored, and I will use this rubric to highlight some of the talks in these areas. For example, omics and novel disease models were widely represented here. Some organelles, like my favorite, the lipid droplet, gained notoriety, and some diseases like PFIC are likely to gain therapeutic traction in the foreseeable future, just based on the number of investigations underway. Omics is definitely the name of the game in 2022. From proteomics to lipidomics, metabolomics, transcriptomics, and spatial transcriptomics, epigenomics, and microbiomics. These technologies enable us to derive biology using advanced computational methodologies. We learned that antibiotics can alter the hepatic transcriptome. Using a CCL4 mouse model of liver injury, Dr. Shannon Samen found that antibiotics not only shift the gut microbiome, but also the hepatic transcriptome, but only in certain pathways, specifically in xenobiotic metabolism and amino acid degradation. While investigators can use global transcriptomics to understand biology, and certainly that is still an important method to use, a major theme here was using single cell omics technology to understand how each cell type contributes to biology. Here, Dr. Ramachandran et al. presented work on the transcriptional signature, not only of these global changes, but also on changes that are happening specifically in endothelial cells. And by using this strategy, they noted that there are several clusters of endothelial cells, in fact. And what's most striking to me is that this was all happening in the cirrhotic scar itself, probably an area of the liver we really don't think about very much. They explored these cell-to-cell interactions that happen specifically in the fibrotic niche itself, and elaborated on the biology of these so-called scar-associated endothelial cells, mesenchymal cells, and macrophages. We learned further from Dr. Chang that applying new visualization tools, in this case highly multiplexed 2D imaging mass cytometry, a mouthful, can advance our understanding of immune complexity. So it's not only helping us understand who is at play, but where they are at play. Combining single-cell transcriptomics with spatial transcriptomics gives us further insight into the where. Technology that layers data derived from single-cell transcriptomics with spatial information, including zonal information, helps us find these so-called hot spots of activity in disease states. Here, Dr. Andrews et al. compared control livers with PSC livers, and in the first panel found a number of differentially expressed genes, especially in the B-cell signatures. And by applying spatial transcriptomics, they found that while the phenotype of B-cells in the two conditions is quite similar, the location of those B-cells differ, and that's really important to understand. In the healthy liver, B-cells tend to be dispersed, while in the PSC liver, they are clustered. Not only can transcriptome analyses help us define disease states, metabolome analyses can as well. Here, Maras et al. help us understand that kidney injury in ACLF patients can take several forms, and these can be distinguished using combinations of metabolites. They looked at patients who didn't have kidney disease and those who had several types of kidney disease phenotype as assessed clinically, and then applied untargeted metabolomics to these individuals, and were able to separate out disease types based on either the up or down regulation of these metabolites. So, for example, in individuals who had the HRS phenotype, they were able to distinguish them by using only five metabolites. But no matter the technology, whether we're talking about transcriptomics or metabolomics, we learned in several presentations here that location, location, location matters. No longer can we simply expect to lump like tissues together to understand biology. Here, for example, for microbiome analyses, Bloom et al. described that in cirrhosis, the duodenum microbiome differs from that of the ileum and differs from that of the colon, and the microbiome of the ileum and colon are closer than that of the duodenum, and also that the duodenal microbiome has lower alpha diversity. There are many microbiome talks and presentations, and they were all fantastically prepared and presented. I only have opportunity to highlight a few today, and these are the ones that led, that really looked at therapeutic strategies for disease intervention. In the upper left-hand corner, this group used a strain of bacteria to ameliorate NASH. Just under that, there was a pooled strain of, a pooled therapeutic strategy of 119 strains. They call it HCOM-2, and this strategy was really employed to hopefully get away from needing to transplant all of the feces from one individual to another and really isolating those that are most promising therapeutically. And the third figure demonstrates perhaps the opposite, that transplantation of fecal EVs can exacerbate ALD. But ultimately, it was Dr. Rehrman's talk that said, wait a minute, let's all pause. Why? Because her group went out into the wild, collected wild mice, brought them back to the lab, and asked the question, are wild mice the same as lab-derived mice, and how are their microbiome similar or different? And indeed, she found that they don't, they aren't the same, that wild mice and lab mice do differ in their gut microbiome. And so we better be sure, as scientists, to have the right model of intervention before moving these trials into humans. The next big theme was genetics, and here I include epigenetics. An important reminder in this meeting, and I saw this in many posters in the poster hall, was that our conclusions are only as good as the data source used for genetic discovery. Here, Rutledge and team incorporated findings from a biobank highly enriched with Latino and Hispanic populations, and found that those with Ecuadorian and Mexican ancestry had the highest allele frequency of the PNPLA3 risk polymorphism, but the lowest frequency of the protective variant of HSD17. They also had the highest liver enzymes, as well as FIB4 scores. These authors teach us that the many so-called undiagnosed or cryptogenic conditions can be diagnosed if we use novel strategies. Here they employ whole exome sequencing. And in this pilot study of 19 patients, they found 26% were given new diagnoses based on sequencing, including two new cases of PFIC. This type of exploration was then used to explain the unknown causes of portal hypertension in this family tree. For the index patient highlighted in the orange box, they found that there was a deficiency of a protein GMAP5. They then went to the preclinical models and asked, well, where is GMAP5 expressed and is it differentially expressed? And indeed they found that there was aberrant expression of this protein in liver sinusoidal endothelial cells and perhaps because of where they see this deficiency and aberrancy, this might explain why these individuals have portal hypertension. Using GWAS studies, we can not only identify individual genes of interest, but explore how these genes work together to confer disease risk. Here is Xiaoming from the Speliotes group established NAFLD subtypes based on polygenic risk scores for PRS. Ascribing names such as low lipoprotein high out or high lipoprotein in to define the epidemiology, for example. And using this strategy, they were able to understand that the higher PRS scores also conferred a higher risk of HCC. PheWAS studies are another approach to gain insight into human polymorphisms. Here Dr. Scorletti from my lab investigates a recently reported human polymorphism of the perilypin-2 lipid droplet protein. While she did not find any association with clinical NAFLD, she did find an association with reduced hepatic triglyceride by approximately 47% reduction. She then took this into the preclinical model and fed mice a high caloric diet. These mice that expressed the human polymorphism had smaller lipid droplets compared to the control mice. And this perhaps explains the difference in their triglyceride content. Finally, in this section of genetics and epigenetics, we learned that it is the yin and yang of genetics and epigenetics that ultimately determine disease. And perhaps surprisingly, epigenetics exhibits plasticity and cell-specific differences. Here I present work of Jasmine from Dr. Thomas' group showing that with HCV infection, interferon-induced isogylation post-translational modifications are regulated by SOX1. And SOX1 activity is dependent on its own methylation status. For this next section, I'll dive into the disease models that were presented at this meeting. Be it the application of omics or genetic tools, our field cannot advance without advances in these preclinical models. And here I will highlight but just a few in vitro and in vivo models that have been developed for liver disease investigation. In in vitro systems, organoids loomed large this meeting, while for in vivo models, dietary and genetic models took center stage. I start with this poster by Yuqin Liu from the Harvard Dental School. First, as a reminder to all of us that hepatology has no departmental borders. These authors are working to improve upon current organoid models, which for the most part exhibits an immature hepatocyte phenotype. Using a maturation cocktail of sorts, they demonstrate that their mature hepatocyte-derived organoid model can be used for investigations in both NAFLD as well as in ALD, and it outperforms the original organoid models. The other work that I will highlight today reflects that the liver is not just its cells. It's also the architecture that enables those cells to interact. Here I highlight a poster by Chen, who is looking at scaffolds, and also I highlight work by Dr. Takebes from his Hans Popper talk, showing how a decellularized scaffold or integration of sinusoidal structure, respectively, can augment organoid models. Chen showed that scaffolds can support ductal morphology, and Takebes showed that they can recreate sinusoids in a dish, and this is, of course, very critical to understanding, ultimately, liver biology. Besides using mice from the wild, as I talked about earlier, this meeting introduced many new models of liver injury. On this slide, I'm only highlighting some titles of interest. One of my favorites is a new gerbil model to study the pathogenesis of chronic hepatitis E. I'm also highlighting two posters from Dr. Kabanda's group, looking at a Pempti knockout mice to study advanced liver injury, and a poster looking at the Gubra-Amlin mouse model of diet-induced and biopsy-confirmed NASH. Combination dietary models were very well represented at the meeting, and here from our group, Dr. Dempsey presented data of a phenocopy of early liver injury due to combined obesogenic and alcohol. Mice were fed a diet high in fructose, fat, and cholesterol with or without daily 5% ethanol plus weekly ethanol gavage for 12 weeks. Performed histology and a combination of lipidomics as well as transcriptomics. Histology demonstrated that the FFC plus alcohol diet was worse than either diet alone, which is what we see in patients who suffer from both insults to the liver. And he found that of the individual components, whether the FFC diet or alcohol, it was the FFC component that conferred the most changes in the transcriptome, affecting pathways such as cholesterol and bile acid metabolism, energy and fatty acid metabolism, and even xenobiotic metabolism and oxidative stress. This model that I'm highlighting here from the Weiman group extends the combination diets to include diphtheria toxin mediated KC cell ablation as a new model that recapitulates several of the features of ALK-HEP. In this model, they see severe steatohepatitis, pericellular fibrosis, and a gene signature that's very similar to the progenitor cell state of ALK-HEP. And finally, I'll highlight a model that can be adapted for therapeutics in ALD and perhaps in other conditions. Here, Lee from the Arteel lab received this poster of distinction for using matrix-bound nano vesicles engineered from healthy livers to treat experimental ALD, with noteworthy improvements in liver histology as well as in markers of liver injury. I'm about to get off of my literal and figurative soapbox, but not before sharing with you with what I think are some real paradigm shifts within liver disease progression understanding. Centuries ago, we understood liver disease as one of the four bad humors, a concept originated by the ancient Egyptians and adapted later by the Greeks. And over time, we made advances in physical exam to help us categorize patients in terms of liver disease severity. We layered on findings with serology and standard histology, H&E, and trichrome. But now I would submit to you that those strategies are insufficiently capturing what we currently understand about liver disease progression. In 2022, it is clear to me that we are now ready to phenotype liver disease based on biology at the intracellular and tissue levels. And in this way, we have the greatest potential to advance our quest for new therapeutics. For example, it is not enough to simply say a patient has steatosis and know what that means in terms of risk for insulin resistance. Dr. Shulman reminds us that it matters how and where these bioactive lipids are sequestered. For example, if they are safely stored in lipid droplets and have limited access to the plasma membrane where they can impair insulin signaling pathways and other important biological pathways, it is unlikely that they will develop insulin resistance. And there are a number of studies really showing that to us now. Here he has depicted this for the example of diacylglycerol species. Where I wanna leave you is perhaps one of the most profound talks that I got to see and learn about this meeting. And it really has shifted my own thinking about how cirrhosis and the complications of cirrhosis develop. It's not simply a progression of histologic scar. I believe that's an antiquated way to think about it in 2022. And here's the evidence that supports it. Campbell's team explains that in fact, these nodules derive from clonal expansion of mutated hepatocytes undergoing distinct selection pressures. It's Darwin at its finest. In the healthy liver, these livers are usually tightly knit and have polyclonal structures. And there may be some clones, but only a few to a thousand at best. But in the cirrhotic liver, there can be thousands of these clones that account for grams of liver tissue, and they tend to respect the nodular architecture. Some of these clones are even monoclonal in nature. And so of course, this brings up questions for hepatocellular carcinoma risk. In an individual patient, one gene might be hit multiple times in multiple cells, and in another patient, it could be a number of different genes. And so that brings up the question of selective pressure, and it's not going to be one size fits all. Here again, the lipid droplet as an organelle is elevated as one of the lipid droplet genes. Side B is one of the most mutated genes in these clonal expansions. FOXO1, a gene that is required for normal insulin signaling is also on this short list. And so for me, it brings up questions of why when patients develop insulin resistance and they have liver disease, it perhaps is such an ominous sign. It also makes me think that we have to move more upstream in our screening and detection and management of patients, because once they have these mutations, it's probably going to be a lot harder to reverse disease or reduce their risk of hepatocellular carcinoma development. So my key takeaways from this meeting are that the liver meeting 2022 showcased many novel hypotheses, advanced technologies, new disease models, and translational promise. I personally found it to be so highly collaborative. I'm optimistic about our prospects for both common and rare disease eradication. And importantly, there are so many opportunities for our emerging investigators to take us to the next level. Thank you for your time. Enjoy the rest of the meeting and safe travels home.

Liver Meeting 2022

hepatology

omics technologies

disease models

gene expression

therapeutics